Software Development on Dislocation Creep in Alloys

Many components made of alloys are subjected to enormous mechanical and thermal loads. Usually, these components can be tested prior to their application and thus be constructed in a safe and cost-efficient way. Unfortunately, there is one phenomenon, which is very difficult to assess experimentally: creep of alloys. Creep is the slow irreversible deformation of a material under high temperature and mechanical loads. Creep can be a significant problem, for example for steam pipes and boilers in thermal power plants or for turbines. Sooner or later the deformation gets too severe for safe operation, and the components can literally explode. This is not only expensive due to repairs and disruption of production, but can endanger the life of people. Since creep deformation is extremely slow, lab tests can take many years, and the development and improvement of materials becomes tedious. As a consequence, currently used standard materials in thermal power plants have been introduced in the 1950s 80s. The materials work but they are inefficient. With new developments, the working temperatures could be raised, leading to higher efficiency, less fuel consumption and lower CO2 emission. The topic of the project is the modelling and simulation of creep. As soon as creep deformation can be simulated with sufficient accuracy, lab tests can be cancelled for the better part and simulations are faster. However, there are two problems. First: creep is extremely complex and difficult to handle. Second: most common models on creep are unreliable and work only under significant confinements; usually a lab test is inevitable (again) in order to tune the simulation towards reality. Nevertheless, there are many good models for partial aspects of the creep process. This is the point where the project sets in: these sub-models are collected and translated into a uniform language. This helps identifying contradicting theories, find gaps and replace them, and combine the bits and pieces into one single network. Each contribution must be physically justified in order to guarantee the predictability of the network. The model network is then coded as glass box software for computer simulations. With this software, users can then easily apply the state of the art (plus the progress achieved within the project) without digging into literature by themselves. The software is then made available for the scientific community and will contribute to accelerate the scientific progress in creep modelling. The development of new materials can be carried out faster, the safety of component can be estimated with better reliability, and technological solutions for higher energy efficiency can be achieved sooner. At the end of the day, the project safes time, money, resources, energy, improves safety and counteracts the greenhouse effect.

Metals and alloys behave fundamentally differently at high temperatures and/or under changing mechanical stress - they gradually deteriorate in their properties until they finally fail. This process can take many years and is therefore very difficult to study experimentally. Simulating the behavior is just as difficult, as the causes can be found deep within the microstructure of the material. The microstructure comprises everything that takes place in the order of magnitude between individual atoms (less than one nanometer) and the visible range (greater than 1/10 millimeter). In general terms, the microstructure includes all deviations of the material from an ideal crystal - since metals are also composed of crystals. Scientifically, it is extremely difficult to predict the service life of a material under these side conditions. Within this project, we took look at the extent to which we can understand this behavior applying the current state of physical models. It was particularly important to us not only to describe the material behavior, but to actually understand it - a huge difference, and an important basis for the development of improved materials. As a result of the project, we obtained a viable model that re-linked known physical principles and improved the individual sub-models in key respects. We were able to close many gaps in the theory and apply one and the same description to a large number of alloy types. As the developed model is extremely complex, we were also looking for a way to simplify the application for other research groups or industrial users. As part of this project, we developed an easy-to-use software tool, into which the model was integrated. The user enters the known initial state of the material and the software simulates the reaction of the material to mechanical stress and high temperatures using the complex model. The main advantages over conventional models are the computing speed (a prediction for many years takes just a few seconds) and the solid physical foundation. If a new material concept, which would first have to be tested over many years, does not show the expected properties in the simulation, the short calculation times allow the cause of the material failure to be investigated quickly. This means that the material can be optimized before it has even been manufactured. This approach will save material developers a lot of time and money in the future.